[ Agenda | Sessions | Program ]

Transitional Dynamics in Non-Scale Growth Models

Theo Eicher and Stephen J. Turnovsky - University of Washington


New growth theory has stressed endogenous knowledge, factor endowments, policy, and tastes as crucial determinants of economic growth. Two key features of these new growth models have been subject to empirical and theoretical criticism. First, both R&D and investment-based growth models that incorporate non-rivalry and/or externalities exhibit scale effects, meaning that variations in the size or scale of the economy permanently alter its long-run equilibrium growth rate. For example, R&D-based growth models that follow Romer's seminal work imply that a doubling of the population and resources devoted to R&D will increase the growth rate proportionately. Investment-based growth models imply that variations in the investment growth rate should permanently influence the growth rate.

The empirical evidence does not support the presence of scale effects. For example, OECD data suggest that variations in the level of research employment exerted no influence on the long-run growth rates of the OECD economies, in contrast with the predictions of the Romer model. In addition, the systematic empirical analysis by Backus, Kehoe, and Kehoe finds no conclusive evidence of a relation between U.S. growth and measures of scale.

A second limitation of recent endogenous growth models is the requirement that to generate an equilibrium of ongoing growth, all production functions must in general exhibit constant returns to scale in the factors of production that are being accumulated endogenously. This strong requirement imposes a strict knife edge restriction on the production structure and has been extensively criticized. For all other degrees of returns to scale, if the economy does converge to a balanced growth equilibrium, it is one that is characterized by the absence of scale effects. From this standpoint, non-scale growth equilibria should be viewed as being the norm, rather than the exception, and consequently it is even more important that their nature be clearly characterized.

In recent work we have analyzed the general structure of a non-scale two-sector growth model, providing a systematic characterization of circumstances that give rise to non-scale growth when key aspects of the economy such as capital and technology are endogenous. That analysis was restricted to balanced growth paths. In this paper we extend that characterization to analyze the transitional dynamics in this new class of endogenous growth model. We begin by presenting the simplest one-sector non-scale model, in which the dynamics can be characterized analytically and compared to both the traditional neoclassical Solow model, as well as to the more recent AK growth model. Nexxt, we investigate the dynami behavior of a two-sector R&D based endogenous growth model. While we are able to characterize its dynamic structure analytically, more explicit analysis requires numerical solutions. Most of our paper focuses on this aspect and in investigating the transitional dynamics of capital and knowledge in response to perturbations of certain key variables.


Scheduled for Session 5.5 Economic Growth

[ Agenda | Sessions | Program ]